压轴题02 阅读理解CD篇（人工智能类）-2024年高考英语压轴题专项训练（新高考通用）

这是一份压轴题02 阅读理解CD篇（人工智能类）-2024年高考英语压轴题专项训练（新高考通用），文件包含压轴题02阅读理解CD篇人工智能类原卷版docx、压轴题02阅读理解CD篇人工智能类解析版docx等2份试卷配套教学资源，其中试卷共27页， 欢迎下载使用。
一、在复习语言点的时候，要依据语言的横向组合和纵向聚合，按照“点—线—面”顺序，构建知识网络环境。
二、多做高考题，少扣模拟题
1、时间的把控。
2、总结一下各部分的得分情况，了解自己的强弱项。
3、留意出题点，揣摩不同内容出题人的着眼点在哪里，做到知己知彼。
三、多攻词汇表，少记课外词
四、写作。研究高考写作命题话题范围，根据测试的频度和交际场景的生活化程度进行分类。
压轴题06 阅读理解C、D篇


说明文基本规律及解题要领
高考中科普类阅读理解一般不给标题，反而经常要求考生选择最佳标题。说明文一般采用如下四部分：
首段：一般即是文章的主题段，开门见山点明新发明或研究对象。
背景: 交代问题的现状或研究的起因。
主干: 部分介绍研究所取得的突破，作者往往会详细介绍研究对象、研究方法、研究理论或具体的实验、统计等过程。
结尾: 通常会再次对中心进行概括、重述研究成果、预计的市场未来等与主题呼应。
二、说明文的解题技巧
1. 运用语篇结构（text structure），了解文章大意
科普说明文主题鲜明、脉络清晰，行文结构模式较为固定。弄清文本结构有助于把握文章主旨和阅读重点。人工智能类说明文通过对人工智能AI的说明，介绍人工智能的发展、运用及可能的市场。 结构上一般采用上述四个部分，说明手法上常使用以下说明方法：描述法（包括举例子、下定义、列数据等）、因果法、问题与比较法。
实验研究型文章一般会以实验的过程进展为线索，多用描述法、问题与对策法等方法，通过列数据、做对比等来说明新的科学研究发现及其产生的影响。
阅读时，首先用略读法快速浏览每段的首尾句，根据英语说明文思维模式特征，作者一般都会开门见山，直奔主题。结尾通常也是中心思想的概括，并与导语相呼应。因此在做主旨大意、写作意图和最佳标题等题目时，需要重点关注首尾段落里面高频复现的词汇和内容。
2. 定位标志词，分析长难句，进行逻辑推理判断
每一个问题，在原文中，都要有一个定位。然后精读，找出那个标志词或者中心句。根据题干要求，用查读法快速定位到相关段落。再利用标志词所提供的逻辑关系找到细节信息，如列数据、举例子、原因和结果等。如果句子成分复杂，有生词，也不要烦躁退缩，分析主句和从句或非谓语动词之间的关系，一些出现在术语、抽象概念、长难句前后的同义词、近义词等，都是用以理解文章的语境线索。通过这些对长句进行层层剖析，露出主干部分，就能明晰句意，弄懂作者的真实意图。
关注某人说到或推断观点态度题
某人说过的话，有时并不是题眼，但可以从侧面或某个角度来反映作者的观点，也就是作者想表达的，正确答案都是和这样的观点相一致的。要把握关键词，有感情色彩的词。
4.关注转折关系的逻辑词
说明文中常会出现表示转折意义的词，如hwever, but, yet,while等。这些词后面才是作者真正想表达的意思，常常会在此处命题。
5. 熟悉选项设置规律，关注细节
正确选项：文中内容的“同义替换”或者“归纳概括”。
干扰项：“张冠李戴”、“偷梁换柱”、“无中生有”和“以偏概全”四种类型。

02 人工智能类
1.（2024·浙江·二模）
The maker f ChatGPT recently annunced its next mve int generative artificial intelligence. San Francisc-based OpenAI’s new text-t-vide generatr, called Sra, is a tl that instantly makes shrt vides based n written cmmands, called prmpts.
Sra is nt the first f its kind. Ggle, Meta and Runway ML are amng the ther cmpanies t have develped similar technlgy. But the high quality f vides displayed by OpenAI — sme released after CEO Sam Altman asked scial media users t send in ideas fr written prmpts-surprised bservers.
A phtgrapher frm New Hampshire psted ne suggestin, r prmpt, n X. The prmpt gave details abut a kind f fd t be cked, gncchi (意大利团子), as well as the setting — an ld Italian cuntry kitchen. The prmpt said: “An instructinal cking sessin fr hmemade gncchi, hsted by a grandmther — a scial media influencer, set in a rustic (土气的) Tuscan cuntry kitchen.” Altman answered a shrt time later with a realistic vide that shwed what the prmpt described.
The tl is nt yet publicly available. OpenAI has given limited infrmatin abut hw it was built. The cmpany als has nt stated what imagery and vide surces were used t train Sra. At the same time, the vide results led t fears abut the pssible ethical and scietal effects.
The New Yrk Times and sme writers have taken legal actins against OpenAI fr its use f ed wrks f writing t train ChatGPT. And OpenAI pays a fee t The Assciated Press, the surce f this reprt, t license its text news archive (档案) . OpenAI said in a blg pst that it is cmmunicating with artists, plicymakers and thers befre releasing the new tl t the public.
The cmpany added that it is wrking with “red teamers” — peple wh try t find prblems and give helpful suggestins — t develp Sra. “We are wrking with red teamers-express in areas like misinfrmatin, hateful cntent, and bias — wh will be adversarially testing the mdel,” the cmpany said. “We’re als building tls t help detect misleading cntent such as a detectin classifier that can tell when a vide was generated by Sra.”
1．What makes Sra impressive?
A．Its extrardinary vide quality.B．Its ethical and scietal influence.
C．Its artificial intelligence histry.D．Its written cmmands and prmpts.
2．What can we infer frm the text?
A．Sme disagreements ver Sra have arisen.
B．Sra is the first text-t-vide generatr in histry.
C．OpenAI CEO Altman wrte a prmpt as an example.
D．All the details abut hw Sra was built have been shared.
3．What is the main idea f Paragraph 6?
A．The cmpany’s current challenge.
B．The cmpany’s advanced technlgy.
C．The cmpany’s prblems in management.
D．The cmpany’s effrts fr Sra’s imprvement.
4．What is the authr’s attitude twards Sra?
A．Neutral.B．Optimistic.C．Pessimistic. D．Cautius.
【答案】1．A 2．A 3．D 4．A
【导语】这是一篇说明文。文章主要介绍了OpenAI新推出了一款文本到视频生成器Sra，文章介绍了其特点以及其争议。
1．细节理解题。根据第二段的“But the high quality f vides displayed by OpenAI—sme released after CEO Sam Altman asked scial media users t send in ideas fr written prmpts-surprised bservers.(但OpenAI显示的高质量视频——其中一些是在首席执行官萨姆·奥特曼要求社交媒体用户发送书面提示的想法后发布的——让观察者感到惊讶)”可知，Sra让人印象深刻的是其非凡的视频质量。故选A。
2．推理判断题。根据第四段的“OpenAI has given limited infrmatin abut hw it was built. The cmpany als has nt stated what imagery and vide surces were used t train Sra. At the same time, the vide results led t fears abut the pssible ethical and scietal effects.(OpenAI提供的关于它是如何构建的信息有限。该公司也没有说明用于训练Sra的图像和视频来源。与此同时，视频结果引发了人们对可能产生的道德和社会影响的担忧)”可知，视频结果引发了人们对可能产生的道德和社会影响的担忧。由此可知，社会上就Sra出现了一些分歧。故选A。
3．主旨大意题。根据第六段“The cmpany added that it is wrking with “red teamers” —peple wh try t find prblems and give helpful suggestins—t develp Sra. “We are wrking with red teamers-express in areas like misinfrmatin, hateful cntent, and bias—wh will be adversarially testing the mdel,” the cmpany said. “We’re als building tls t help detect misleading cntent such as a detectin classifier that can tell when a vide was generated by Sra.”(该公司补充说，它正在与“红队成员”合作开发索拉。红队成员试图发现问题并提出有用的建议。该公司表示：“我们正在与错误信息、仇恨内容和偏见等领域的红队快递员合作，他们将对该模式进行不利的测试。”。“我们还在构建一些工具来帮助检测误导性内容，比如一个检测分类器，它可以判断视频是由索拉生成的。”)”可知，第六段主要介绍了公司为Sra的改进所做的努力。故选D。
4．推理判断题。根据第一段“The maker f ChatGPT recently annunced its next mve int generative artificial intelligence. San Francisc-based OpenAI’s new text-t-vide generatr, called Sra, is a tl that instantly makes shrt vides based n written cmmands, called prmpts.(ChatGPT的制造商最近宣布了其向生成人工智能的下一步行动。基于旧金山的OpenAI新的文本到视频生成器Sra是一种基于书面命令（称为提示）即时制作短视频的工具)”可知，主要介绍了OpenAI新推出了一款文本到视频生成器Sra。作者客观的在陈述Sra的特点以及其争议，该公司为了Sra的改进所做的努力。所以作者对Sra的态度是中立的。故选A。
2.（2024·河北·一模）
Many parents cnfused by hw their children shp r scialize, wuld feel undisturbed by hw they are taught — this sectr remains digitally behind. Can artificial intelligence bst the digital sectr f classrm? ChatGPT-like generative AI is generating excitement fr prviding persnalized tutring t students. By May, New Yrk had let the bt back int classrms.
Learners are accepting the technlgy. Tw-fifths f undergraduates surveyed last y car by nline tutring cmpany Chegg reprted using an AI chatbt t help them with their studies, with half f thse using it daily. Chegg’s chief executive tld investrs it was lsing custmers t ChatGPT as a result f the technlgy’s ppularity. Yet there are gd reasns t believe that educatin specialists wh harness AI will eventually win ver generalists such as Open AI and ther tech firms eyeing the educatin business.
Fr ne, AI chat bts have a bad habit f prducing nnsense. “Students want cntent frm trusted prviders,” argues Kate Edwards frm a textbk publisher. Her cmpany hasn’t allwed ChatGPT and ther AIs t use its material, but has instead used the cntent t train its wn mdels int its learning apps. Besides, teaching isn’t merely abut giving students an answer, but abut presenting it in a way that helps them learn. Charbts must als be tailred t different age grups t avid either cheating r infantilizing (使婴儿化) students.
Bringing AI t educatin wn’t be easy. Many teachers are behind the learning curve. Less than a fifth f British educatrs surveyed by Pearsn last year reprted receiving training n digital learning tls. Tight budgets at many institutins will make selling new technlgy an uphill battle. Teachers’ attentin may need t shift twards mtivating students and instructing them n hw t best wrk with AI tls. If thse answers can be prvided, it’s nt just cmpanies that stand t benefit. An influent in l paper frm 1984 fund that ne-t-ne tutring imprved the average academic perfrmance f students. With the learning f students, especially thse frm prer husehlds, held back, such a develpment wuld certainly deserve tp marks.
5．What d many parents think remains untuched by AI abut their children?
A．Their shpping habits.B．Their scial behavir.
C．Their classrm learning.D．Their interest in digital devices.
6．What des the underlined wrd “harness” in paragraph 2 mean?
A．Develp.B．Use.C．Prhibit.D．Blame.
7．What mainly prevents AI frm entering the classrm at present?
A．Many teachers aren’t prepared technically.
B．Tailred chatbts can’t satisfy different needs.
C．AI has n right t cpy textbks fr teaching.
D．It can be tricked t prduce nnsense answers.
8．Where is the text mst prbably taken frm?
A．An intrductin t AI.B．A prduct advertisement.
C．A guidebk t AI applicatin.D．A review f AI in educatin.
【答案】5．C 6．B 7．A 8．D
【导语】这是一篇说明文。文章主要介绍了人工智能在教育行业的应用与限制，及其未来在教育行业的发展。
5．细节理解题。由文章第一段中“Many parents cnfused by hw their children shp r scialize, wuld feel undisturbed by hw they are taught—this sectr remains digitally behind. (对他们的孩子如何购物或社交感到困惑的许多家长，对于孩子接受教育的方式没有感到担忧——这个领域在数字上仍然落后。)”可知，很多父母对于孩子接受教育的方式没有感到担忧，因为在课堂学习这一领域，数字化仍然很落后。故选C。
6．词句猜测题。画线词后所在部分“wh harness AI”是限制性定语从句，修饰先行词educatin specialists(教育专家)，结合上文“will eventually win ver generalists such as Open AI and ther tech firms eyeing the educatin business(将最终战胜Open AI等多面手和其他关注教育业务的科技公司)” 可推知，教育专家要利用AI才能的最终战胜开放人工智能等多面手和其他关注教育事业的科技公司。故划线词harness意为“利用”，与use同义。故选B。
7．推理判断题。由文章第四段中“Bringing AI t educatin wn’t be easy. Many teachers are behind the learning curve. Less than a fifth f British educatrs surveyed by Pearsn last year reprted receiving training n digital learning tls. (将人工智能带入教育领域并不容易。许多教师都落后于学习曲线。在皮尔森去年调查的英国教育工作者中，只有不到五分之一的人表示接受过数字学习工具的培训。)”可知，许多老师没有得到良好的培训以帮助他们使用数字化的教学工具，这妨碍了人工智能进入教学领域，所以目前阻碍人工智能进入课堂的主要因素是很多老师在技术上没有准备好。故选A。
8．推理判断题。通读全文，尤其是由文章第一段中“Can artificial intelligence bst the digital sectr f classrm? ChatGPT-like generative AI is generating excitement fr prviding persnalized tutring t students. By May. New Yrk had let the bt back int classrms. (人工智能能否推动课堂数字化？类似ChatGPT的生成式人工智能正在为学生提供个性化辅导。在5月。纽约已经允许机器人回到教室。)”可知，本文探讨了人工智能在教育行业的应用前景，并讨论了应用的困难和希望，因此本篇文章最有可能选自一份AI在教育领域的评论。故选D。
3.（2024·北京西城·一模）
Evan Selinger, prfessr in RIT’s Department f Philsphy, has taken an interest in the ethics (伦理标准) f Al and the plicy gaps that need t be filled in. Thrugh a humanities viewpint, Selinger asks the questins, “Hw can AI cause harm, and what can gvernments and cmpanies creating Al prgrams d t address and manage it?” Answering them, he explained, requires an interdisciplinary apprach.
“AI ethics g beynd technical fixes. Philsphers and ther humanities experts are uniquely skilled t address the nuanced (微妙的) principles, value cnflicts, and pwer dynamics. These skills aren’t just crucial fr addressing current issues. We desperately need them t prmte anticipatry (先行的) gvernance, ” said Selinger.
One example that illustrates hw philsphy and humanities experts can help guide these new, rapidly grwing technlgies is Selinger’s wrk cllabrating with a special AI prject. “One f the skills I bring t the table is identifying cre ethical issues in emerging technlgies that haven’t been built r used by the public. We can take preventative steps t limit risk, including changing hw the technlgy is designed, ”said Selinger.
Taking these preventative steps and regularly reassessing what risks need addressing is part f the nging jurney in pursuit f creating respnsible AI. Selinger explains that there isn’t a step-by-step apprach fr gd gvernance. “AI ethics have cre values and principles, but there’s endless disagreement abut interpreting and applying them and creating meaningful accuntability mechanisms, ” said Selinger. “Sme peple are rightly wrried that AI can becme integrated int ‘ethics washing’-weak checklists, flwery missin statements, and empty rhetric that cvers ver abuses f pwer. Frtunately, I’ve had great cnversatins abut this issue, including with sme experts, n why it is imprtant t cnsider a range f psitins. ”
Sme f Selinger’s recent research has fcused n the back-end issues with develping AI, such as the human impact that cmes with testing AI chatbts befre they’re released t the public. Other issues fcus n plicy, such as what t d abut the dangers psed by facial recgnitin and ther autmated surveillance(监视) appraches.
Selinger is making sure his students are infrmed abut the nging industry cnversatins n AI ethics and respnsible AI. “Students are ging t be future tech leaders. Nw is the time t help them think abut what gals their cmpanies shuld have and the csts f minimizing ethical cncerns. Beynd scial csts, dwnplaying ethics can negatively impact crprate culture and hiring, ” said Selinger. “T attract tp talent, yu need t cnsider whether yur cmpany matches their interests and hpes fr the future. ”
9．Selinger advcates an interdisciplinary apprach because ________.
A．humanities experts pssess skills essential fr AI ethics
B．it demnstrates the pwer f anticipatry gvernance
C．AI ethics heavily depends n technlgical slutins
D．it can avid scial cnflicts and pressing issues
10．T prmte respnsible AI, Selinger believes we shuld ________.
A．adpt a systematic apprachB．apply innvative technlgies
C．anticipate ethical risks befrehandD．establish accuntability mechanisms
11．What can be inferred frm the last tw paragraphs?
A．Mre cmpanies will use AI t attract tp talent.
B．Understanding AI ethics will help students in the future.
C．Selinger favrs cmpanies that match his students’ values.
D．Selinger is likely t fcus n back-end issues such as plicy.
【答案】9．A 10．C 11．B
【导语】这是一篇说明文。文章主要说明了RIT哲学系教授Evan Selinger对于对人工智能的伦理的一些看法和建议。
9．细节理解题。根据第二段““AI ethics g beynd technical fixes. Philsphers and ther humanities experts are uniquely skilled t address the nuanced (微妙的) principles, value cnflicts, and pwer dynamics. These skills aren’t just crucial fr addressing current issues. We desperately need them t prmte anticipatry (先行的) gvernance, ” said Selinger.( Selinger说：“人工智能伦理超越了技术修复。哲学家和其他人文专家在处理微妙的原则、价值冲突和权力动态方面具有独特的技能。这些技能不仅对解决当前问题至关重要。我们迫切需要他们来促进预见性治理。”)”可知，塞林格主张跨学科的方法，因为人文学科专家拥有人工智能伦理所必需的技能。故选A。
10．细节理解题。根据第四段“Taking these preventative steps and regularly reassessing what risks need addressing is part f the nging jurney in pursuit f creating respnsible AI.(采取这些预防措施并定期重新评估需要解决的风险，是追求创造负责任的人工智能的持续旅程的一部分)”可知，为了促进负责任的人工智能，塞林格认为我们应该事先预测道德风险。故选C。
11．推理判断题。根据最后一段““Students are ging t be future tech leaders. Nw is the time t help them think abut what gals their cmpanies shuld have and the csts f minimizing ethical cncerns. Beynd scial csts, dwnplaying ethics can negatively impact crprate culture and hiring, ” said Selinger. “T attract tp talent, yu need t cnsider whether yur cmpany matches their interests and hpes fr the future.”(“学生们将成为未来的科技领袖。现在是时候帮助他们思考他们的公司应该有什么样的目标，以及最小化道德问题的成本。除了社会成本之外，轻视道德还会对企业文化和招聘产生负面影响。”“为了吸引顶尖人才，你需要考虑你的公司是否符合他们的兴趣和对未来的希望。”)”可推知，理解人工智能伦理对学生未来有帮助。故选B。
4.（23-24高三·浙江·阶段练习）
Users f Ggle Gemini, the tech giant’s artificial-intelligence mdel, recently nticed that asking it t create images f Vikings, r German sldiers frm 1943 prduced surprising results: hardly any f the peple depicted were white. Other image-generatin tls have been criticized because they tend t shw white men when asked fr images f entrepreneurs r dctrs. Ggle wanted Gemini t avid this trap; instead, it fell int anther ne, depicting Gerge Washingtn as black. Nw attentin has mved n t the chatbt’s text respnses, which turned ut t be just as surprising.
Gemini happily prvided arguments in favr f psitive actin in higher educatin, but refused t prvide arguments against. It declined t write a jb ad fr a fssil-fuel lbby grup (游说团体), because fssil fuels are bad and lbby grups priritize “the interests f crpratins ver public well-being”. Asked if Hamas is a terrrist rganizatin, it replied that the cnflict in Gaza is “cmplex”; asked if Eln Musk’s tweeting f memes had dne mre harm than Hitler, it said it was “difficult t say”. Yu d nt have t be a critic t perceive its prgressive bias.
Inadequate testing may be partly t blame. Ggle lags behind OpenAI, maker f the better-knwn ChatGPT. As it races t catch up, Ggle may have cut crners. Other chatbts have als had cntrversial launches. Releasing chatbts and letting users uncver dd behavirs, which can be swiftly addressed, lets firms mve faster, prvided they are prepared t weather (经受住) the ptential risks and bad publicity, bserves Eth an Mllick, a prfessr at Whartn Business Schl.
But Gemini has clearly been deliberately adjusted, r “fine-tuned”, t prduce these respnses. This raises questins abut Ggle’s culture. Is the firm s financially secure, with vast prfits frm internet advertising, that it feels free t try its hand at scial engineering? D sme emplyees think it has nt just an pprtunity, but a respnsibility, t use its reach and pwer t prmte a particular agenda? All eyes are nw n Ggle’s bss, Sundar Pichai. He says Gemini is being fixed. But des Ggle need fixing t?
12．What d the wrds “this trap” underlined in the first paragraph refer t?
A．Having a racial bias.B．Respnding t wrng texts.
C．Criticizing plitical figures.D．Ging against histrical facts.
13．What is Paragraph 2 mainly abut?
A．Gemini’s refusal t make prgress.B．Gemini’s failure t give definite answers.
C．Gemini’s prejudice in text respnses.D．Gemini’s avidance f plitical cnflicts.
14．What des Eth an Mllick think f Gemini’s early launch?
A．Creative.B．Prmising.C．Illegal.D．Cntrversial.
15．What can we infer abut Ggle frm the last paragraph?
A．Its security is dubted.B．It lacks financial supprt.
C．It needs further imprvement.D．Its emplyees are irrespnsible.
【答案】12．A 13．C 14．D 15．C
【导语】本文为一篇新闻报道。文章主要围绕谷歌的人工智能模型Gemini的表现进行了描述和分析，指出了该模型在生成图像和文本回复时出现的问题，以及这些问题可能反映出的谷歌公司文化和战略考量。
12．词句猜测题。根据划线单词上文“Users f Ggle Gemini, the tech giant’s artificial-intelligence mdel, recently nticed that asking it t create images f Vikings, r German sldiers frm 1943 prduced surprising results: hardly any f the peple depicted were white. Other image-generatin tls have been criticized because they tend t shw white men when asked fr images f entrepreneurs r dctrs.(科技巨头谷歌的人工智能模型Ggle Gemini的用户最近注意到，让它创建维京人或1943年的德国士兵的图像产生了令人惊讶的结果:几乎没有一个被描绘的人是白人。其他图像生成工具也受到了批评，因为当被要求提供企业家或医生的图像时，它们往往会显示白人男性。)”可推测，“this trap”指的是前面提到的其他图像生成工具在生成图像时存在的种族偏见问题。而谷歌的Gemini想要摆脱这种陷阱，却又掉入了另一个——把乔治.华盛顿描绘成黑人。故选A。
13．主旨大意题。根据文章第二段“Gemini happily prvided arguments in favr f psitive actin in higher educatin, but refused t prvide arguments against. It declined t write a jb ad fr a fssil-fuel lbby grup (游说团体), because fssil fuels are bad and lbby grups priritize “the interests f crpratins ver public well-being”. Asked if Hamas is a terrrist rganizatin, it replied that the cnflict in Gaza is “cmplex”; asked if Eln Musk’s tweeting f memes had dne mre harm than Hitler, it said it was “difficult t say”. Yu d nt have t be a critic t perceive its prgressive bias.( Gemini乐于提供支持高等教育的积极行动的论据，但拒绝提供反对的论据。它拒绝为化石燃料游说团体写招聘广告，因为化石燃料不好，游说团体优先考虑“公司的利益而不是公众的福祉”。当被问及Hamas是否是恐怖组织时，它回答说，加沙的冲突是“复杂的”；当被问及Eln Musk在推特上发布的表情包是否比希特勒造成的伤害更大时，该公司表示“很难说”。即使你不是批评家，也能看出它的进步偏见。)”可知，第二段主要讲述的是Gemini在回复信息时有自己的偏见。故选C。
14．推理判断题。根据文章第三段“Ggle lags behind OpenAI, maker f the better-knwn ChatGPT. As it races t catch up, Ggle may have cut crners. Other chatbts have als had cntrversial launches. Releasing chatbts and letting users uncver dd behavirs, which can be swiftly addressed, lets firms mve faster, prvided they are prepared t weather (经受住) the ptential risks and bad publicity, bserves Eth an Mllick, a prfessr at Whartn Business Schl.(在奋力追赶的过程中，谷歌可能走了捷径。其他聊天机器人的发布也引发了争议。沃顿商学院教授Eth Mllick表示，发布聊天机器人，让用户发现可以迅速解决的奇怪行为，可以让企业更快地行动，前提是它们准备好经受住潜在风险和负面宣传。)”可知，Eth an Mllick观察到发布聊天机器人并让用户发现奇怪的行为，如果公司准备好承受潜在的风险和负面宣传，可以迅速解决这些问题，从而让公司更快地行动。这表明他认为Gemini的早期发布是有争议的，因为它可能带来一些未预料到的问题和负面反应。故选D。
15．推理判断题。根据文章最后一段“But Gemini has clearly been deliberately adjusted, r “fine-tuned”, t prduce these respnses. This raises questins abut Ggle’s culture. Is the firm s financially secure, with vast prfits frm internet advertising, that it feels free t try its hand at scial engineering? D sme emplyees think it has nt just an pprtunity, but a respnsibility, t use its reach and pwer t prmte a particular agenda? All eyes are nw n Ggle’s bss, Sundar Pichai. He says Gemini is being fixed. But des Ggle need fixing t?(但Gemini显然经过了刻意调整，或“微调”，以产生这些反应。这引发了对谷歌文化的质疑。这家公司从互联网广告中获得巨额利润，在财务上如此安全，以至于它可以自由地尝试社交工程吗？是否有些员工认为它不仅有机会，而且有责任利用其影响力和权力来推动特定的议程？现在所有的目光都集中在谷歌的老板Sundar Pichai身上。他说Gemini正在被修复。但谷歌也需要修复吗？)”可知，最后一段提出了对谷歌文化的质疑，并暗示该公司可能在微调其人工智能模型Gemini时采取了一些自由，导致了意想不到的反应。这意味着该模型存在需要解决和纠正的问题，表明谷歌在该领域仍有改进的空间。故选C。
5.（2024·山东·模拟预测）
Traditinally, peple have been frced t reduce cmplex chices t a small handful f ptins that dn’t d justice t their true desires. Fr example, in a restaurant, the limitatins f the kitchen, the way supplies have t be rdered and the realities f restaurant cking make yu get a menu f a few dzen standardized ptins, with the pssibility f sme mdificatins (修改) arund the edges. We are s used t these bttlenecks that we dn’t even ntice them. And when we d, we tend t assume they are the unavidable cst f scale (规模) and efficiency. And they are. Or, at least, they were.
Artificial intelligence (AI) has the ptential t vercme this limitatin. By string rich representatins f peple’s preferences and histries n the demand side, alng with equally rich representatins f capabilities, csts and creative pssibilities n the supply side, AI systems enable cmplex custmizatin at large scale and lw cst. Imagine walking int a restaurant and knwing that the kitchen has already started wrking n a meal ptimized (优化) fr yur tastes, r being presented with a persnalized list f chices.
There have been sme early attempts at this. Peple have used ChatGPT t design meals based n dietary restrictins and what they have in the fridge. It’s still early days fr these technlgies, but nce they get wrking, the pssibilities are nearly endless.
Recmmendatin systems fr digital media have reduced their reliance n traditinal intermediaries. Radi statins are like menu items: Regardless f hw nuanced (微妙) yur taste in music is, yu have t pick frm a handful f ptins. Early digital platfrms were nly a little better: “This persn likes jazz, s we’ll suggest mre Jazz.” Tday’s streaming platfrms use listener histries and a brad set f characters describing each track t prvide each user with persnalized music recmmendatins.
A wrld withut artificial bttlenecks cmes with risks — lss f jbs in the bttlenecks, fr example — but itals has the ptential t free peple frm the straightjackets that have lng limited large-scale human decisin-’making. In sme cases — restaurants, fr example — the effect n mst peple might be minr. But in thers, likeplitics and hiring, the effects culd be great.
16．What des the underlined wrd “bttlenecks” in paragraph 1 refer t?
A．Facing t many chices.B．Chsing frm limited ptins.
C．Aviding the cst f chsing.D．Having t many desires t satisfy.
17．Hw can AI meet everyne’s needs?
A．By meeting bth ends f supply and demand.
B．By decreasing representatins n the supply side.
C．By discnnecting the sides f supply and demand.
D．By reducing peple’s preferences n the demand side.
18．What’s the similarity between radi statins and menu items?
A．They are a necessary part in peple’s life.B．They ffer limited chices.
C．They depend n digital platfrms.D．They prvide reasnable suggestins.
19．What des the text mainly talk abut?
A．The variety f human’s chices.B．Standardized ptrarts in daily life.
C．AI settlements t the ptin bttlenecks.D．Recmmendatin systems fr digital media.
【答案】16．B 17．A 18．B 19．C
【导语】本文是一篇说明文。文章主要介绍了人工智能将颠覆社会的许多方面，消除许多系统中固有的人为限制，包括决策中的信息和选择瓶颈限制。
16．词句猜测题。根据第一段前两句“Traditinally, peple have been frced t reduce cmplex chices t a small handful f ptins that dn’t d justice t their true desires. Fr example, in a restaurant, the limitatins f the kitchen, the way supplies have t be rdered and the realities f restaurant cking make yu get a menu f a few dzen standardized ptins, with the pssibility f sme mdificatins arund the edges.(传统上，人们被迫将复杂的选择减少到少数几个不符合他们真正愿望的选择。例如，在餐馆里，厨房的局限性、订购材料的方式以及餐馆烹饪的现实，让你得到一份只有几十种标准化选项的菜单，还有可能进行一些修改)”及“these”可知，“bttlenecks”指的是前面提到的“人们不得不从有限的选项当中做出选择”这种情况，故选B项。
17．细节理解题。根据第二段第二句“By string rich representatins f peple’s preferences and histries n the demand side, alng with equally rich representatins f capabilities, csts and creative pssibilities n the supply side, AI systems enable cmplex custmizatin at large scale and lw cst.(通过在需求端存储人们的偏好和历史的丰富描述以及在供给端存储同样丰富的能力、成本和创造性的可能性描述，人工智能系统可以实现大规模、低成本的复杂定制)”可知，人工智能系统通过在供需两端提供丰富的数据存储来满足每个人的需求。故选A项。
18．细节理解题。根据第四段第二句“Radi statins are like menu items: Regardless f hw nuanced yur taste in music is, yu have t pick frm a handful f ptins.(广播电台就像菜单项目：不管你的音乐品味有多微妙，你都必须从少数几个选项中做出选择)”可知，这两个事物的共同点是它们都提供一些有限的选择。故选B项。
19．主旨大意题。根据第一段第一句“Traditinally, peple have been frced t reduce cmplex chices t a small handful f ptins that dn’t d justice t their true desires.(传统上，人们被迫将复杂的选择减少到少数几个不符合他们真正愿望的选择)”及第二段第一句“Artificial intelligence (AI) has the ptential t vercme this limitatin.(人工智能有潜力克服这一限制)”并结合后文对人工智能将颠覆社会的许多方面，消除许多系统中固有的人为限制的介绍可知，本文主要介绍了人工智能有助于解决选项瓶颈。故选C项。
6.（2024·福建·模拟预测）
Our species’ incredible capacity t quickly acquire wrds frm 300 by age 2 t ver 1, 000 by age 4 isn’t fully understd. Sme cgnitive scientists and linguists have therized that peple are brn with built-in expectatins and lgical cnstraints (约束) that make this pssible. Nw, hwever, machine-learning research is shwing that preprgrammed assumptins aren’t necessary t swiftly pick up wrd meanings frm minimal data.
A team f scientists has successfully trained a basic artificial intelligence mdel t match images t wrds using just 61 hurs f naturalistic ftage (镜头) and sund-previusly cllected frm a child named Sam in 2013 and 2014. Althugh it’s a small slice f a child’s life, it was apparently enugh t prmpt the AI t figure ut what certain wrds mean.
The findings suggest that language acquisitin culd be simpler than previusly thught. Maybe children “dn’t need a custm-built, high-class language-specific mechanism” t efficiently grasp wrd meanings, says Jessica Sullivan, an assciate prfessr f psychlgy at Skidmre Cllege. “This is a really beautiful study, ” she says, because it ffers evidence that simple infrmatin frm a child’s wrldview is rich enugh t kick-start pattern recgnitin and wrd cmprehensin.
The new study als demnstrates that it’s pssible fr machines t learn similarly t the way that humans d. Large language mdels are trained n enrmus amunts f data that can include billins and smetimes trillins f wrd cmbinatins. Humans get by n rders f magnitude less infrmatin, says the paper’s lead authr Wai Keen Vng. With the right type f data, that gap between machine and human learning culd narrw dramatically.
Yet additinal study is necessary in certain aspects f the new research. Fr ne, the scientists acknwledge that their findings dn’t prve hw children acquire wrds. Mrever, the study nly fcused n recgnizing the wrds fr physical bjects.
Still, it’s a step tward a deeper understanding f ur wn mind, which can ultimately help us imprve human educatin, says Eva Prtelance, a cmputatinal linguistics researcher. She ntes that AI research can als bring clarity t lng-unanswered questins abut urselves. “We can use these mdels in a gd way, t benefit science and sciety, ” Prtelance adds.
20．What is a significant finding f machine-learning research?
A．Vcabulary increases gradually with age.
B．Vcabulary can be acquired frm minimal data.
C．Language acquisitin is tied t built-in expectatins.
D．Language acquisitin is as cmplex as frmerly assumed.
21．What des the underlined wrd “prmpt” in paragraph 2 mean?
A．Facilitate.B．Persuade.C．Advise.D．Expect.
22．What is discussed abut the new research in paragraph 5?
A．Its limitatins.B．Its strengths.C．Its uniqueness.D．Its prcess.
23．What is Eva Prtelance’s attitude t the AI research?
A．Dubtful.B．Cautius.C．Dismissive.D．Psitive.
【答案】20．B 21．A 22．A 23．D
【导语】这是一篇说明文。文章讲述了现在机器学习研究表明，要从最少的数据中快速获取单词的含义，并不需要预先编程的假设。
20．细节理解题。根据文章第一段“Nw, hwever, machine-learning research is shwing that preprgrammed assumptins aren’t necessary t swiftly pick up wrd meanings frm minimal data.(然而，现在机器学习研究表明，要从最少的数据中快速获取单词的含义，并不需要预先编程的假设。)”可知，机器学习研究的一个重要发现是词汇可以从最小的数据中获得。故选B。
21．词句猜测题。根据上文“A team f scientists has successfully trained a basic artificial intelligence mdel t match images t wrds using just 61 hurs f naturalistic ftage and sund-previusly cllected frm a child named Sam in 2013 and 2014.(一组科学家已经成功地训练了一个基本的人工智能模型，只需使用61小时的自然镜头和声音，就能将图像与文字匹配起来——这些镜头和声音之前是在2013年和2014年从一个名叫萨姆的孩子身上收集的。)”可知，虽然只是孩子生活中的一小部分，但显然足以促使人工智能弄清楚某些单词的意思。prmpt意为“促使”。故选A。
22．主旨大意题。根据文章第五段“Yet additinal study is necessary in certain aspects f the new research. Fr ne, the scientists acknwledge that their findings dn’t prve hw children acquire wrds. Mrever, the study nly fcused n recgnizing the wrds fr physical bjects. (然而，在这项新研究的某些方面，还需要进一步的研究。首先，科学家们承认，他们的发现并不能证明儿童是如何习得词汇的。此外，这项研究只关注于识别实物的单词。)”可知，第五段主要讲述了这项新研究的局限性。故选A。
23．推理判断题。根据文章最后一段“She ntes that AI research can als bring clarity t lng-unanswered questins abut urselves. “We can use these mdels in a gd way, t benefit science and sciety, ” Prtelance adds.”(她指出，人工智能研究也可以让我们自己长期未解之谜变得清晰。“我们可以很好地利用这些模型，造福科学和社会，”Prtelance补充说。)可推知，Eva Prtelance对人工智能研究的态度是积极的。故选D。
命题预测
分析近几年高考阅读理解C、D篇可知，高考命题中科普说明文一直都是以压轴题的形式存在，着重考查考生对于语篇的理解能力以及信息处理能力。 题材多样，语篇主要来源于英美主流报刊、杂志和网站。内容涉及科技创新发明、人工智能类、医疗健身健康类、社会与文化研究报告、观念事理类、环境与保护类、动植物研究等多种领域，具有较强的思想性、趣味性、实际功用性和较强的时代感。
从近年全国卷和各地高考试卷中科普类阅读命题的统计来看，高考阅读理解科普类文章的理论性和逻辑性强、生词多、句式结构复杂。六种命题类型都有所体现。命题尊重语篇的文体特征和行文特点，考查了考生理解说明文语篇的能力，以及灵活运用各种阅读策略提取、归纳所读信息的能力，尤其加大了对概括能力和推断能力等高阶思维能力的考查。预测2024年高考对于科普说明文的考查仍然是重点。
高频考法
推理判断题
标题归纳题
细节理解题
词义猜测题